With the development of radio communication technologies and the growing popularity of mobile intelligent terminals, more and more users choose to access the Internet through mobile terminals, to participate in activities such as office working, learning or entertainments, so that there are a wide variety of mobile data services. At present, 2G-based mobile data services are generally based on general packet radio service (GPRS) technologies. The core network of a mobile data service of 3G communication is also evolved on the basis of the GPRS. A gateway GPRS support node (GGSN) is a network element of a packet switch (PS) domain in the core network (CN) of the mobile network, and provides for data packets a gateway interface function between the GPRS network and an external data network. The deep packet identify (DPI) of the GGSN network element is able to perform deep identification on all the packet messages that are being transmitted so as to parse out information such as a service type. A policy and charging control (PCC) system based on the DPI technology provides flexible management of and control over control policies and charging policies for the mobile data services. In the system, a user may subscribe to a corresponding control policy package in a policy and charging rule function (PCRF). The control policy and the charging policy of the user and the service are issued by the PCRF based on a session state and subscription information. The corresponding control and charging are executed by the policy and charging enforcement function (PCEF) based on the policies issued from the PCRF. However, for mobile data network applications, the policy and charging control mechanism cannot dynamically adjust the control policies of the user according to information such as the current network state, the user's behaviour characteristics and the usage condition of the service to reversely control the mobile network; therefore, the following defects exist in the related art:
(1) when the mobile network has a high load, it is impossible to intelligently protect very important persons (VIP) or important and high timeliness mobile data services from influences;
(2) when the mobile network has a low load, it is impossible to intelligently improve the quality of service (QoS) of the network in the user policy, maximize the throughput of the network and improve the network experiences of the user; and
(3) it is impossible to intelligently protect the mobile network load so as to avoid overload.
Regarding at least one of the above-mentioned problems in the related art, no effective solution has been proposed.